The Design Of The Fiber Automatic Recognition Software System Based On Image Processing Technology

The fiber identifies to analyze the composition of the textile product has the great significance for quantity evaluation and function analysis of the fiber product.In the thesis the fiber images processing technologies is studied and a fiber image automatic recognition system is developed. The system can perform online fiber image automatic recognition which can overcome the disadvantages of the fiber image recognition by manual way. The acquisition of the fiber image is implemented by the portable photoelectric instrument for fiber inspection, then the images are transmitted into PC through the bus interface of USB. After the image preprocessing ,five feature such as energy entropy of the fiber was extracted based on the gray level co-occurrence matrix. Finally, based on the feature paraments, the classification of the nearest neighbor was chosen for the recognition.In the thesis an integrated software system is submitted, which combined image processing and analyzing using a visual programming software, namely, Visual C++ 6.0. A feature database that can save and manage the feature values of fiber image was also included in the system.